Listen Reader:
an electronically augmented
paper-based book
Maribeth Back, Jonathan Cohen, Rich Gold, Steve Harrison, Scott Minneman
Xerox PARC
3333 Coyote Hill Road
Palo Alto, CA USA
+1 650 812 4726
back, harrison, minneman, richgold@parc.xerox.com; cohen@dnai.com
ABSTRACT
While predictions abound that electronic books will
supplant traditional paper-based books, many people
bemoan the coming loss of the book as cultural artifact. In
this project we deliberately keep the affordances of paper
books while adding electronic augmentation. The Listen
Reader combines the look and feel of a real book — a
beautiful binding, paper pages and printed images and text
— with the rich, evocative quality of a movie soundtrack.
The book's multi-layered interactive soundtrack consists of
music and sound effects. Electric field sensors located in
the book binding sense the proximity of the reader's hands
and control audio parameters, while RFID tags embedded
in each page allow fast, robust page identification.
Three different Listen Readers were built as part of a sixmonth museum exhibit, with more than 350,000 visitors.
This paper discusses design, implementation, and lessons
learned through the iterative design process, observation,
and visitor interviews.
Keywords
Audio books, interactive museum exhibits, electronic
books, augmented reality, augmented books, multimodal
i/o, new genres, sound design, interactive audio,
interactive books, page detection, RFID tags, embedded
tags, gestural input, smart documents.
PROBLEM STATEMENT: ELECTRONIC BOOKS
Electronic books have been the focus of an enormous
amount of research, both commercial and academic. Prior
art ranges from Alan Kay’s DynaBook to the current crop
of commercial CD-ROM publishers and PDA and e-book
manufacturers. None, however, have taken the approach of
enabling complex electronic interaction through the
physical use of a real book. These electronic books are
generally presented as an onscreen computer application,
as in a CD-ROM, or as a handheld device with a computer
screen. The electronic book is almost never conceived as a
real book, with real paper pages. Even in the case of
"electric paper" the so-called paper is unlikely to offer the
look and feel of paper constructed of wood chips, nor will
the book much resemble the traditional bound book.
Currently, electronically augmented books tend towards
three models:
a) button-based interactive children's
storybooks, with embedded audio buttons; b) online or CDROM books and storybooks; and c) the handheld e-book or
PDA-style electronic reader, generally designed for
downloadable texts. (As of this writing, a real e-paper
book is not available.)
The first instance, where the electronic book is a real book,
is children's audio storybooks, which often have short
sounds associated with particular images. The sound in
these interactive books, in particular those with buttons
built into the pages or bindings, is simplistic and often
distracts the reader rather than enhancing or illustrating
the reading material. However, nearly all of them are
designed in a fashion that interrupts the reader’s attention
to the story by forcing a search for the “hot spots” in the
images or text -- the trigger points for the sound, which
requires the activation of a button or switch to play and
pause. In addition, the sound itself is usually very poor
quality, played in short, harsh bursts. The frequency
content of such bursts is such that it is difficult for the
human perceptual system to successfully put them in the
background; this too can disrupt the reader’s involvement
with the story.
Online or CD-ROM-based "books" lack many of the
attributes of books. They retain a book-like depth of
content and sometimes a book design interaction
metaphor, using pages and chapters, for example; and
publishers of CD-ROM books talk about their product as
books. But these "books" must be read on a computer, and
the distinction between a CD-ROM or online "book" and
other computer applications is unclear. In this case the
term "electronic book" is stretched thin, and clearly has
more to do with the structure of content than with form or
physical affordances.
The e-book or PDA handheld device model of the
electronic book has gained a great hold on the imagination
of the publishing industry. Though several companies are
now marketing these devices, the e-book model of the
future of the book has spurred many complaints. [8] In
sum: e-books are expensive hardware, they are not
sufficiently robust, they have poor resolution or contrast or
font formation or there is some other difficulty which
makes the screen hard to read for very long. The interface
is too present; it requires too much attention; downloading
is too difficult and requires extra equipment. Readers find
themselves aware of the device, rather than the content;
thus immersive reading is difficult if not impossible.
Though many of these issues should ease with time and
developments in both technology and social expectations
around reading, some large questions remain: do people
really want to do all their reading on one single device? Is
the form and feel of a book a major affordance in focusing
the attention of the person reading it?
The Listen Reader draws from research in ubiquitous
computing, tangible media, and augmented reality. [3, 11,
12] It pushes back on the concepts of the convergent
reading device (as in e-books), of book-as-application (CDROMs) or book-as-keyboard (button-based audio
storybooks). The Listen Reader is designed to preserve the
cherished experience of immersive reading and to preserve
the beauty of the paper-based book as token object.
Earlier work: the SIT Book
The Listen Reader is the fruition of several years of
research focused on new document genres and the
multimodal possibilities of interactive reading, especially.
We reported on the SIT (Sound-Image-Text) Book, an
early version of this work, in a short paper at CHI'99. [2]
With the SIT Book we were primarily interested in the use
of audio as an added layer of illustration for a children's
book. We also were interested in applying audio as a
peripheral layer, added information that was nice to have
but that was not necessary to understanding the book. In
contrast, the Listen Reader's sounds are an integral part of
the interactive reading experience.
Context: museum setting
The Listen Readers are part of a museum exhibition
featuring eleven interactive experiences designed to
explore the relationship between technology and reading.
The six-month exhibit XFR: Experiments in the Future of
Reading was installed at the Tech Museum of Innovation
in San Jose, California. More than 350,000 visitors came
through the exhibit between March and September 2000.
Because the exhibit took place within the context of a
modern technology museum, the XFR exhibits are
primarily interactive and hands-on, in keeping with many
of the other exhibits.
Fig. 1. The Listen Reader provides a classic immersive
reading environment: a comfortable chair, a polished
hardwood reading stand, and beautiful paper pages in a
soft leather book binder. High-quality embedded audio and
responsive proximity ("magic") sensors add to the sense of
immersion.
Description
The Listen Reader consists of a large comfortable
wingback chair with a small but excellent sound system
embedded in it: a small speaker in each wing, not far from
a person’s ear, and a subwoofer under the seat. A wooden
swing arm table sits to the right of the chair; on this table
is affixed a book with leather and brass binding and handsewn, colorful paper pages. Embedded in the pages and
binding of the book – but invisible to the user – are the
sensors that enable the electronic augmentation of the
book.
The Listen Reader combines an interactive multi-track
audio environment with text and images printed on
traditional paper pages. Ambient sounds added to each
page heighten the drama of the story. As the reader
casually fingers different parts of the page, the sound
gently adapts to the different moods and scenes. The
sound environment is continuous; it stops only when the
reader finishes the book, turning the last page.
Here are some examples of typical interactions with a
Listen Reader:
A museum visitor -- in this case, a 12-year-old
girl -- approaches the three Listen Readers sitting
in a calm corner of the museum. Curious at
finding comfortable stuffed chairs in this hightech environment and attracted by the brightly
colored books, she sinks into one of the chairs and
pulls the book towards her. Suddenly, in response
to her hand on the book's cover, she hears a cat's
purr. The sounds come from the chair she's
sitting in; her head swings around for a moment,
trying to locate the source of the sounds, but then
she is drawn back to the book. As she opens it to
the first page, a rich soundscape fades in: a
burbling stream, buzzing cicadas, chirping birds,
and an excerpt from John Coltrane's version of
"Stairway to the Stars." At this point the girl
realizes she has encountered something new, and
she pauses a moment to read the instructional
signage mounted on the table next to the book.
"Move your hands all around the book – you don't
even have to touch the pages" she is advised, and
as she progresses through the book, reading the
story on each page, she plays with the sounds by
moving her hands, exploring the different
atmospheres she can create. As she turns each
page, the current soundscape crossfades into the
soundscape for the next page; she is never left
sitting in silence, not is there an abrupt, harsh
transition. Typically, she stays to finish the entire
book, and then moves on to read the next Listen
Reader.
Another visitor clearly has had much experience
with the audio-button storybooks; he tries
punching various illustrations to elicit sounds,
and is confused when the response is not an
instant on/off but rather a volume fade-up or
down. Eventually he figures out that he has a
different type of control over this sound, and gives
up on punching the illustrations. Though he never
looks at the instructional signage and does not
figure out how to use all of the interaction, he
stays to read through the book, content with the
shifting soundscapes he gets from simply turning
the pages.
A mother sits down in the Listen Reader and piles
her three kids around her on the arms of the chair
and on her lap. She reads the story aloud to them
as they all play with the sounds in the book.
Happy to have found a group experience for all of
them (and relieved to be off their feet) they read
the book all the way through.
MULTIMODAL INTERACTION DESIGN
Listen Reader is an experiment in multimodal reading. In
addition to rich graphics and imaginative text, the
modality of sound is woven into the storyline. We are
interested in sound as well as graphics as part of a
multimodal symbol set for reading. Sound as well as
graphics can offer strongly affective imagery. Sound also
serves to establish a sense of place. Thus, sound can
operate as its own additional stylized symbol system, to be
read along with the graphics and text. [4]
In the Listen Reader, another point of the audio is to
enhance the feeling of communing with a private world,
one that gently acknowledges one’s own presence. This is
similar to the sense of immersion one feels in reading a
good book. Combining the two provides an even more
immersive experience.
Continuous control versus buttons
An important feature of the Listen Reader is its design for
integrating the audio illustration into the story. Interactive
children's books often encourage the reader to turn to a
new page and immediately punch every image to see
whether it makes a noise. We move away from the binary
button methodology by giving the reader the soundscape
for each page immediately, and then providing the reader
with continuous control mechanisms. The task is no longer
to find the sounds; it is rather to mix them is a satisfying
way, as an accompaniment to the story.
The continuous controllers were mapped to relate the
volume of a sound to the proximity of a hand (or other
body part). As the reader’s hand approaches a certain area
of the page, the sound gets louder. It fades when the hand
is removed. The continuous controller is tuned to react at
the same apparent rate of speed as the hand’s motion. The
continuous interaction is less likely to induce the seek-andpunch behaviors common in button-based systems. We
feel that this allows an important shift in focus for the
interactive reader.
SOUND DESIGN AND CONTENT
We chose to begin work with childrens’ books, a type of
literature where images and text tend to have an almost
equal relationship. The three Listen Readers were all
children's books from San Francisco's Chronicle Books, for
which we designed unique soundtracks. The books were
Hipcat, by Jonathan London, illustrated by Woodleigh
Hubbard; Armadillo Ray, by John Beifuss, illustrated by
Peggy Turley (we used the Spanish language version,
translated by Augustin Antreasayan); and Frank was a
Monster Who Wanted To Dance, by Keith Graves. Music
for HipCat was from several John Coltrane albums, used
with permission from Atlantic Records; we designed the
sound effects.
Each book had a relatively small amount of text per page,
and full-page color illustrations ranging from natural
settings (deserts, streams, the ocean) to depictions of
characters and activities (dancing, eating, driving, reading
cat poetry, applauding). The sounds available on each page
were related to the images found there. When the page is
opened, all the sounds for that page begin to run softly in a
continuous loop. There are four sounds associated with
each page, one in each quadrant, and each loop is of a
different length; thus the effect is that of a changing sonic
landscape, not a repeating sound loop.
We chose a running background ambience so that the
reader would never fall out of the immersive atmosphere
created by the sound. Silence, especially abrupt silence
upon a page turn, seems to require the visitor to do
something. In the Listen Readers, focus is designed to
remain on the story and the multimodal reading
experience, rather than on the interaction.
SYSTEM DESIGN
The Listen Reader system combines two technologies:
embedded RFID tagging for page identification, and
electric field sensing to read proximity data from the hands
of the person(s) reading the book. From these two data
streams we can establish which sets of sounds are to be
played, and what kind of manipulation should be
performed on each sound (usually volume or pitch control,
or both).
Page identification
Systems that gather accurate page-ID data from paperbased books have generally involved either mechanical
switch-based systems that are prone to false reads or
optical systems that have particular lighting or visual
requirements. Though many systems have been attempted,
none are in common use.
Camera-based page recognition systems require line-ofsight placement and particular lighting conditions. There
are also page-ID systems that require the reader to perform
a specific act, such as press a button, run a pen over a
barcode, or pass the page through a reader. However these
solutions are not transparent; they interfere with a person’s
natural interaction with the book. For example, Stifelman
showed a paper-based electronic notebook that allowed the
user to take notes on real paper [9]. Those notes could be
cross-indexed electronically because the page-ID was
known; but the user had to make sure the edge of the page
fit into an optical reader along the side of the notebook.
Although it was an elegant solution, it was not completely
transparent to the user.
Our method of page-ID uses a new RFID technology (the
TIRIS TagIt system from Texas Instruments, with
simultaneous ID) to recognize what page a book is open to.
[1, 10] A thin flexible transponder tag with a unique ID is
embedded in the paper of each page, and a special tagreader is affixed to the binding of the back of the book.
As the pages turn, the tag-reader notices which tags are
within its read range and which have moved out of its
range (which is about four inches). The human interacts
with the book naturally, and is not required to perform any
actions that are not usual in book interaction.
The Listen Reader code (written in Director Lingo, with a
specially written serial Xtra) sets the tag-reader so that it
continually checks for a list of tags within its reading
range. This list is received by the computer (WinNT 4.0)
and updated as frequently as possible (approximately once
a second). As a page is turned, the embedded tag leaves
the reading range of the tag-reader, and its absence is
noticed.
Geometry in tag layout on the pages is of some account. In
this system, two tags may overlap by 80% or more, but not
completely. This overlap issue is addressable through the
development of smaller tags (now available) or
multichannel tag readers, or better SID (Simultaneous ID)
algorithms.
Fig. 2. RFID transponders (tags). Flexible copper printed
onto clear plastic, embedded between two printed sheets to
create one page thickness for the Listen Reader books.
Each transponder has a unique ID. As of September 2000,
tags about half this size are available.
We use this information to provide appropriate contextual
information associated with that page in the book: in this
instance, music and sound effects. When a page is turned,
the sounds associated with the first page fade down in
volume while the sounds associated with the new page
fade up in volume.
Other possibilities include spoken word, specific point
lighting, ambient room lighting, moving graphics on a
computer screen, and scrolling text.
Other tag systems
The Motorola Bistatix tag-based ID system is another
example of a thin, flexible tag system that could be used to
determine pageturning in realtime. Unlike the Texas
Instruments RFID system, which is a passive transponder,
the Bistatix is a capacitive field sensing tag, each having
its own rewritable ID. This system, though still in
development, is perhaps even more useful for our
application since it could conceivably be printed directly
onto a paper page, rather than having to be embedded
between two sheets like the TagIt. Paradiso’s sweptfrequency tags are another interesting tag technology that
could be used in this fashion [5].
Research Labs, the QT110. [7] The sensor measures
conductivity within a field. Placing a hand in the sensor's
electromagnetic fields will shunt some of the current to
ground. The sensor detects the slight current drop, and can
track the hand as it passes through the field in all three
dimensions. The sensor does not provide a solid x-y-z set
of coordinates, as the variables are many (size and shape of
person's arm/hand, muscle density, angle of approach,
whether they're wearing any metal, etc.). However, it does
provide a good real-time set of data that mirrors a user’s
activity accurately.
In this configuration there are four electrodes in the book
binding, one in each of the four corners. This
configuration establishes an electromagnetic field
corresponding to each quadrant of the book. The sensor is
detecting and tracking the reader’s hand as it passes
through each of the four quadrants.
Physical design
Antenna in binding
RFID
reader
A/D
converter
with
serial output
WINNT
PC w/
Director
Fig. 3. The RFID antenna is built onto the binding behind
right-hand page. As the page is turned, its transponder
moves out of the read range of the antenna; the software
notices and switches priority to the next page beneath.
Proximity sensing mapped to control parameters
The other technology used in the Listen Readers is electric
field sensing. We use it as proximity-sensing technology,
embedded in what is essentially a smart binder for the
pages of a book. Early versions of this work used an eightchannel sensor called the LazyFish that was developed at
the MIT Media Laboratory. [6] The Listen Reader uses a
similar commercially available sensor from Quantum
Because of the museum setting, with its huge number of
expected visitors (more than 350,000) the Listen Readers
were required to be extremely robust, both electronically
and in physical design. The pages were a non-tear paper
and the book signatures were hand-sewn with a highstrength kite string. The book was attached to a swing arm
table, which, along with the chair, was bolted down to a
heavy steel plate. All electronics were hidden.
We also wanted to encourage some specific behaviors
associated with reading: sitting and spending some time
with a book, for example. So, the Listen Readers were
designed to provide the look and feel of the comfortable
reading corner in one's living room: a cozy chair, good
but intimate lighting, beautiful hardwood furniture. The
books themselves were hand-bound with brass and leather,
and printed on high-quality color printers. The audio was
stereo CD-quality, and care was taken to tune each system
for premium frequency response (taking into consideration
that the speakers were chair-mounted). By creating a
“classy” look and feel, we encouraged visitors to treat the
Listen Readers with greater care than if the chairs had
been encased in vinyl and the books bolted down with
laminated pages.
RESULTS
We were frankly surprised at some of our observations on
the Listen Readers, both technical and ideological. Other
observations bore out or expanded our original
suppositions. The reports below were based on our daily
repair and maintenance reports and on observations by our
group and museum staff.
Repair, remediation and redesign
Some adjustments were made over the course of the sixmonth-long exhibition. We developed better systems for
replacing the books and their associated files and for
rebooting the machines. We fine-tuned the sound content
and adjusted the sensitivity of the sensors. As we expected,
we had to reset the volume levels to take into account the
noisy environment of an active museum. To begin with, we
were afraid that the proximity of the three Listen Readers
to each other (about fifteen feet apart, set in a triangle)
would create cacaphony, with audio from each interfering
with the others. However, due to the focus and placement
of the speakers within the wings and seat of the armchair,
the sound was much louder for the person in the chair than
for people standing around watching. In this way the
sound added to the personal immersive experience
provided by the Listen Reader: a sort of "sound sphere"
surrounded people in the chair, an audio atmosphere
which rapidly dropped off even two feet away.
We found several different "failure modes" as follows: a)
all hardware fine, software frozen; b) software fine, RFID
hardware system fine, but the A/D card for the proximitysensing hardware frozen; or c) all hardware and software
apparently fine but the system making no noise. Some
occurred once or twice a week, and others occurred only
once or twice total (no sound, or RFID frozen). We quickly
learned the different types of disk cleanups, reboots and
power cycles that should be applied in each case, and
designed a script-based daily reboot system that addressed
most of these conditions satisfactorily. By midway
through the exhibition's run, maintenance on the Listen
Readers was routine. In general, they worked; the Listen
Readers (like the rest of the XFR exhibit) had less
downtime than most of the other museum exhibits.
Robustness
One pleasant surprise was the robust nature of the Listen
Readers. We worried that the wooden swing-arm tables
might crack under the weight of children leaning on them;
that the books would be the target of graffiti artists; that
the books were too fragile to withstand daily handling by
hundreds of museum goers and that we'd need to replace
them every few days. (It turned out be about once every
three weeks.) We expected to have to replace or at least
launder the heavy cotton chair covers; but over the course
of six months, they withstood the daily onslaught without
trauma.
In fact, none of our envisioned disaster scenarios occurred.
We believe that this is in part due to the "living room"
look and feel of the Listen Readers: beautiful wood, highquality colorful printed books with real leather binders,
soft pooled lighting that highlighted the chair and the
person in it. The design of the exhibit itself invited visitors
to slow down, sink into the cushions, and enter a calmer
state of mind. Our musuem partners also believe that it is
in part due to the general character of the visitors to the
San Jose Tech Museum; some museums in some areas of
the country are notoriously harder on exhibits than others.
Visitor interviews
We observed or interviewed a number of the museum
visitors who visited the XFR show. This was done in four
ways: informally, through watching and videotaping the
visitors' interactions with the Listen Readers; through
conversations with the visitors; more formally in an
observation report done by museum staff; and a set of ten
sit-down interviews with museum visitors who had just
finished using the Listen Readers.
Flaunting museum conventions
Received wisdom in the museum world is that visitors do
not read. About 30 seconds of time is all that most
exhibits can expect to receive, according to the consultants
and museum staff we talked with. However, the average
time spent at a Listen Reader was much longer, at least
several minutes. Frequently visitors would read all the
way through one Listen Reader, and then move on to the
next. We had hoped that we could encourage people to
linger (that's why we built three) but this exceeded our
expectations. We believe this is partly because we
presented reading as the content of the exhibit, not as
signage.
Social reading
One surprise was that the Listen Readers encouraged
social reading among the museum visitors. It was at least
as common to see two or three or four people (usually
children) piled up in one Listen Reader chair as it was to
see a single person. Often one person would read the book
aloud while the others played with the sounds (this was
especially common among family groups, such as a parent
with children).
Understanding the interaction
Some visitors were confused by the proximity sensors,
expecting to find the kind of on/off functionality of buttonbased systems. Some such visitors would figure out the
difference after reading a few pages, watching other
visitors, or reading the instructional signage; others
remained frustrated, wanting a clearly causal relationship
between the images and the sound, rather than hand
motion and sound. We believe that this problem is
addressable through a more fine-grained matrix of control
(more than four sensors per page spread). It's also
addressable through the complete authoring of books for
the Listen Reader system, rather than using existing books
to which we add soundtracks.
FUTURE WORK
The XFR: Experiments in the Future of Reading museum
exhibition, of which the Listen Reader is part, will visit
several other museums in the next few years. The lessons
learned in each new installation will produce even more
robust versions of the Listen Reader's software and
hardware. Content may also be redesigned and rewritten;
one project is to author a multimodal book for the Listen
Reader from start to finish, rather than our current method
of adding a soundtrack to existing books. Authoring a
Listen Reader from start to finish means we can address
the apparent mismatch in causality (what sound is related
to which image?) that a few of our visitors mentioned. We
can make sure that either the content is authored to each
quadrant of the book page, or that the sensors adapt
themselves to understand where the important illustrations
are upon each page.
Versions of the Listen Reader have applications in several
areas other than museum exhibits: for example, language
training, musical training, auditory illustration for the
blind, and multimodal books for those with ADD
(attention deficit disorder). Although the current set of
Listen Readers control audio, the same technology could
easily drive any kind of dynamic content, such as lights,
projections, and so on. For example, a contractor’s book of
blueprints could control a computer projection or a 3D set
of blueprints – by turning to a particular page in a
notebook, the contractor could access appropriate and upto-the-minute drawings, cost estimates, or design samples.
We intend to explore some of these applications. We also
intend to make a portable version of the Listen Reader,
with built-in wireless electronics and a data link to a local
area network.
CONCLUSION
Our model of a paper-based electronic book holds promise.
People understood its use with far greater alacrity than we
had expected, and at least in this application, were
delighted to discover such complex functionality in what
appeared to be a real book.
The Listen Readers were designed with the idea that form
affects meaning, and in fact is inextricable from it. We
found that by authoring the form as conscientiously as the
content, we were able to achieve some unusual goals:
getting people to read deeply in a museum setting, for
example, and getting people to read socially, in groups,
often aloud to each other. By leveraging people's own
knowledge and expectations about the uses of real books,
we were able to augment the affordances of the book form
with those of an electronic information system.
ACKNOWLEDGMENTS
The authors would like to thank our colleagues in the RED
group at Xerox PARC, as well as the San Jose Tech
Museum of Innovation, Atlantic Records, and Chronicle
Books of San Francisco. We would also like to thank
Peggy Szymanski and Victoria Bellotti who aided us with
the visitor interviews, and Terry Murphy of Exhibit
Engineering.
REFERENCES
1. Back, Maribeth, and J. Cohen. "Page Detection Using
Embedded Tags." Proceedings of UIST 2000, ACM Press,
2000.
2. Back, Maribeth, R. Gold and D. Kirsch, “The SIT Book:
Audio as Affective Imagery in Interactive Storybooks.”
Human Factors in Computing Systems; CHI99 Extended
Abstracts, 1999.
3. Ishii, H. and Ullmer, B., Tangible Bits: Towards
Seamless Interfaces between People, Bits and Atoms, in
Proceedings of Conference on Human Factors in
Computing Systems (CHI '97), (Atlanta, March 1997),
ACM Press, pp. 234-241.
4. Mynatt, E., M. Back and R. Want, “Designing Audio
Aura.” Proceedings of CHI ’98, ACM Press, 566-573.
5. Paradiso, Joseph and K. Hsiao, "Swept-Frequency,
Magnetically-Coupled Resonant Tags for Realtime,
Continuous, Multiparameter Control," Human Factors in
Computing Systems; CHI99 Extended Abstracts, 1999.
6. Paradiso, J. and N. Gershenfeld, “Musical Applications
of Electric Field Sensing.” October 1995 Computer Music
Journal). Also jrs.www.media.mit.edu/people/jrs/lazyfish/
7. Quantum Research Group. White paper on proximity
sensors. http://www.qprox.com/whatis.htm
8. Silberman, Steve. “Ex Libris.” Wired, July 1998.
9. Stifelman, Lisa. “Augmenting Real-World Objects: A
Paper-Based Audio Notebook.” Proceedings of CHI’96,
ACM Press.
10. TIRIS. Tag-it Inlays, www.tiris.com, Product Bulletin,
Texas Instruments, 1999.
11. Want, Roy, Ken Fishkin, Beverly Harrison, Anuj
Gujar. "Bridging Real and Virtual Worlds with Electronic
Tags", ACM SIGCHI. May 1999, Pittsburgh, PA.
12. Weiser, M. "The Computer for the 21st Century."
Scientific American, 265(3), 1991, pp. 94-104.